Stop commanding ChatGPT: describe outcomes, constraints and examples for better results
Aiden_Tech_Ai · x · 2026-09-15
A shared prompt-engineering tip: instead of giving commands, tell the model the outcome you want.
- Don't say "write code" — describe the feature, constraints, and example inputs/outputs
- Don't say "fix this error" — share the error message, relevant code, and what you expected to happen
- Treat an advanced AI like a capable collaborator and you'll get better results
The thread promises 7 copy-paste prompt templates (in follow-up posts).
More from Apps
- A PM live-built a Slido alternative with zero coding: "coding is solved, delegate it to agents" — PawelHuryn · 2026-09-15
- Complained about no Siri AI beta invite on X, got one within 15 minutes — film_girl · 2026-09-15
- Voice agent Fo ships major voice stack update: call it, and it calls for you — Scobleizer · 2026-09-15
- Ataraxis Debuts Causal AI That Predicts Cancer Treatment Outcomes Zero-Shot — multiply_matrix · 2026-09-15
- NexDoc launches agent design tool: frontier-quality docs at 53X lower price, 3X faster — amiyapatanaik · 2026-09-15
- Meta's Alexandr Wang Praises Muse App as the Next AI Form Factor — alexandr_wang · 2026-09-15